Protective responses against M. tuberculosis (Mtb) in humans are poorly understood, especially those that would prevent establishment of infection or progression to disease. There is still no efficacious vaccine against Mtb, although ~30 vaccines are in various stages of testing and clinical trials. New treatment and prevention strategies are desperately needed to make a major impact on transmission, morbidity and mortality of TB. The host-pathogen interactions occurring during Mtb infection are complex and span across scales, from bacterial and cellular to organ to an entire host. To address this complex disease we need comprehensive and integrative tools to generate testable hypotheses about what characterizes an effective immune response to Mtb infection. Understanding the immune response to Mtb requires a systems biology approach, particularly a computational tool that can integrate large amounts and different types of data regarding specific aspects of the host-pathogen interaction in TB. This project represents an integrated strategy between computational and experimental approaches to tackle this challenging problem. The pathologic hallmark of Mtb infection is a granuloma, a collection of host cells (e.g. macrophages and T cells) that organize in an attempt to contain or eliminate the infection. Within a single host, several granulomas form in response to initial infection, and these granulomas are heterogeneous with variable trajectories, complicating the study of this infection. T cells play a central role in protection against TB, as best exemplified by the dramatic susceptibility of HIV+ humans to TB, even in the early stages of HIV infection. However, T cells come in many functional sub-types. De-convoluting the combination of T cell phenotypes and function that are most efficacious in clearing infection is a herculean task, one that requires a systems biology approach, marrying relevant experimental models and multi-scale and multi-compartment computational models. In the proposed work, we incorporate new data into a next-generation multi-scale and multi-compartment (lung-blood-lymph) computational model that has a host-scale readout. We pair with human and NHP studies both outlined herein and in ongoing separately funded studies to calibrate and validate the models and use them as a testing ground for model predictions in an iterative fashion in 3 key aims: (1) Characterize experimentally the heterogeneity, specificity, and localization of T cells and their function in granulomas and lymph nodes early post-Mtb infection in NHPs and use these data to parameterize, refine and validate our next-generation computational model adding important mechanisms that are currently lacking. (2) Identify mechanisms that balance pro and anti-inflammatory signals in granulomas and distinguish hypotheses regarding host and bacterial factors that limit granuloma T-cell function. (3) Identify early adaptive responses that prevent establishment of infection or disease using virtual clinical trials. Our established collaboration over 15 years will serve well the aims of this interdisciplinary proposal.